Biocatalysis Assistant

Biocatalysis Assistant is an advanced Python-based chatbot framework designed to automate and facilitate bioinformatics and biocatalysis tasks. Leveraging a dynamic set of tools and integrating powerful language models, this assistant streamlines complex processes in computational biology and enzyme engineering, making research more efficient and accessible.

Features

• 🗣️ Interactive Chatbot Interface: Engage with the assistant through a user-friendly command-line interface.
• 🔧 Dynamic Tool Integration: Seamlessly run various bioinformatics and biocatalysis tools.
• 🤖 Language Model Integration: Utilize state-of-the-art language models to guide analysis and decision-making processes.
• 🔗 Extensible Architecture: Easily integrate new tools and functionalities.
• 📈 Optimization Capabilities: Optimize enzyme sequences for improved catalytic activity.
• 🔍 Comprehensive Analysis: Perform detailed bioinformatics analyses, including binding site extraction and reaction element analysis.

Tools

The Biocatalysis Assistant integrates several powerful tools to assist in various bioinformatics and biocatalysis tasks:

• ExtractBindingSites: Extract binding sites from enzyme sequences.
• GetElementsOfReaction: Analyze and extract elements from biochemical reactions.
• Blastp: Perform BLASTP searches for protein sequence analysis.
• OptimizeEnzymeSequences: Optimize enzyme sequences for improved catalytic activity.
• FindPDBStructure: Search for related protein structures in the PDB database.
• DownloadPDBStructure: Download PDB structure files.
• Mutagenesis: Perform in silico mutagenesis on protein sequences.
• MDSimulation: Run molecular dynamics simulations.

Installation

Prerequisites

• Python: Version 3.10 or higher. Download Python

To use all the supported tools, install the following on your system:

• BLASTP: NCBI BLAST+ toolkit. Download BLASTP
• PyMOL: Molecular visualization system. Install PyMOL via Conda
• GROMACS: Molecular dynamics package. Install GROMACS

Steps

1. Clone the Repository

   git clone [email protected]/GT4SD/lm-assistant-for-biocatalysis
   cd lm-assistant-for-biocatalysis

2. Install Poetry and Project Dependencies

   pip install poetry
   poetry install

3. Setup molecular dynamics file, the rxnaamapper and enzyme optimization models Choose one of the following options based on your operating system:

   a) For macOS users with Homebrew:

   bash tools_setup/setup.sh

   b) For users on other operating systems or without Homebrew:

   bash tools_setup/setup.sh --skip-mc

   The script performs the following actions:

   (i) RXNAAMapper Setup - Sets up thea BERT model, a tokenizer and vocabulary

   (ii) Minio Client (mc) Setup (skipped if using --skip-mc) - Installs Minio Client using Homebrew - Configures Minio Client with predefined credentials

   (iii) Enzyme Optimization Models Setup - Downloads and sets up models for enzyme optimization (Minio needed)

   Note: - If you're not using macOS or don't have Homebrew, use the --skip-mc option and manually install Minio Client appropriate for your operating system.

4. Install Additional Required Tools

   • BLASTP
   • PyMOL
   • GROMACS

5. Configure Environment Variables

   Create a .env file in the project root and add your configuration:

   KEY=your_key_here
   API=your_api_here

   Note: Ensure that .env is added to .gitignore to protect sensitive information.

Configuration

Each tool within the Biocatalysis Assistant has its own configuration options that can be customized using environment variables. Refer to the Tool Documentation for specific configuration details.

Usage

Running the Biocatalysis Assistant

The Biocatalysis Assistant offers two interfaces for interaction: a Command Line Interface (CLI) and a Streamlit web application. Choose the interface that best suits your needs.

Command Line Interface (CLI)

To start the Biocatalysis Assistant using the CLI:

1. Ensure you have completed the installation process using Poetry.
2. Open your terminal.
3. Run the following command:

lmabc --help


Biocatalysis Assistant CLI Help

DESCRIPTION:
    This CLI provides an interface to interact with the Biocatalysis Assistant,
    allowing users to ask questions and receive responses related to biocatalysis.

USAGE:
    lmabc

INITIAL SETUP:
    Upon starting, you will be prompted to:
    1. Set verbosity level
    2. Confirm or change the model and provider

COMMANDS:
    help   : Display this help message
    exit   : Exit the program
    clear  : Clear the screen

INTERACTION:
    After setup, you can start asking questions. The assistant will process your
    input and provide responses based on its knowledge of biocatalysis.

EXAMPLES:
    - To ask a question:
      > Enter your question: What are the main types of enzyme catalysis?

    - To clear the screen:
      > Enter your question: clear

    - To exit the program:
      > Enter your question: exit

Streamlit Web Application

To run the Biocatalysis Assistant using the Streamlit web interface:

1. Ensure you have completed the installation process using Poetry.
2. Open your terminal.
3. Run the following command:

lmabc-app

This command starts the Streamlit server and automatically opens the web application in your default browser. If it doesn't open automatically, the terminal will display a local URL you can copy and paste into your browser.

The Streamlit app provides a user-friendly interface with:

• A chat-like interaction on the home page
• A list of available tools
• Settings for customizing the assistant's configuration

Choose the interface that best fits your workflow and preferences. Both interfaces interact with the same underlying Biocatalysis Assistant, ensuring consistent functionality and results.

References

If you use lmabc in your projects, please consider citing the following:

@software{LMABC,
  author = {Yves Gaetan Nana Teukam, Francesca Grisoni, Matteo Manica},
  month = {10},
  title = {The biocatalysis assistant: a language model agent for biocatalysis (lmabc)},
  url = {https://github.com/GT4SD/lm-assistant-for-biocatalysis},
  version = {main},
  year = {2024}
}

License

The lmabc codebase is under MIT license. For individual model usage, please refer to the model licenses found in the original packages.

Support

For issues or questions, please open an issue in the GitHub repository.